SCRO / Journées de l'optimisation

HEC Montréal, 29-31 mai 2023


HEC Montréal, 29 — 31 mai 2023

Horaire Auteurs Mon horaire

RAHSI Resource Allocation in Healthcare Systems I

29 mai 2023 10h30 – 12h10

Salle: Rona (bleu)

Présidée par Vinicius Martins Ton

4 présentations

  • 10h30 - 10h55

    A ward occupancy forecasting model as a tool to increase elective surgery loads in pandemic recovery settings

    • Mehmet Begen, Ivey Business School
    • Felipe Rodrigues, prés., School of Management, Economics and Mathematics King’s University College at Western University
    • Gregory Zaric, Professor, Management Science, Ivey School of Business, Western University

    In the second semester of 2022, Ontario hospitals were overwhelmed by their elective surgery waitlist. For example, ward utilization in London Health Sciences Centre (LHSC) was at its highest, while a mandate to reduce the surgery wait list was a pandemic recovery priority. We developed a stochastic model to predict the impact of extra elective surgery patients on downstream intensive care unit (ICU), step-down unit (SDU) and ward utilization. The model was implemented in MS Excel and helped managers understand the impact of increased surgery loads on capacity and utilization and improved managers' confidence when planning surgery schedules and resource allocation decisions. Its use and insights have enabled LHSC to confidently increase its elective surgery loads and reduce the size of the waitlist and wait times for patients.

  • 10h55 - 11h20

    Optimizing Hospital Resources for COVID-19 Patient Care: A Discrete Event Simulation Approach to Ward and ICU Bed Planning

    • Daniel Garcia-Vicuña, prés., Public University of Navarre
    • Laida Esparza, University Hospital of Navarre
    • Fermin Mallor, Public University of Navarre

    This study involved developing a simulation model to help hospitals make decisions about how to allocate resources in the short term during outbreaks like the COVID-19 pandemic. The model needed to be able to accurately represent the current state of the system and mimic its dynamics to be effective. It consists of two main components: stochastic modeling of patient admission and patient flow processes. Patient arrival is modeled using a Gompertz growth model, which can represent the exponential growth of the virus, followed by a period of maximum arrival rate and then a decrease until the wave subsides. Patient flow is modeled by considering different pathways and estimating length of stay in different healthcare stages. Empirical studies showed that the Gompertz model was the best fit for pandemic-related data and had better predictive capacity than other sigmoid models. The simulation model was applied in two regions of Spain during the COVID-19 waves of 2020. The model was used daily to inform the regional healthcare planning team, who programmed ward and ICU beds based on the resulting predictions.

  • 11h20 - 11h45

    Solving the Intra-hospital Bimodal Transport Problem

    • Vinicius Martins Ton, prés., Centre interuniversitaire de recherche sur les reseaux d'entreprise, la logistique et le transport (CIRRELT)
    • Angel Ruiz, Centre interuniversitaire de recherche sur les reseaux d'entreprise, la logistique et le transport (CIRRELT)
    • José Eduardo Pécora Jr., Centre interuniversitaire de recherche sur les reseaux d'entreprise, la logistique et le transport (CIRRELT)

    Patient transportation is a crucial hospital operational activity. Therefore, the mismanagement directly impacts the hospital's operations, costs, and patients' perceived level of service. Assuming that a set of transport requests, with their respective transport modes (wheelchairs or beds) and porters, are given in advance, the problem consists of simultaneously selecting the request to perform and the porter to serve it. Handling different transport modes requires switching equipment: a porter must visit a depot to pick up/drop off a wheelchair before/after serving a request. Good management of the resources is crucial for the system's proper operation, thus ensuring the ideal amount of resources to meet the demand.

    Efficient approaches are proposed to tackle this context's dynamic intra-hospital patient multi-modal transportation problem. The aim is to tackle the problem of assignment of transport requests to porters, considering multiple transport modes, and also present an analysis of resource utilization. We model the assignment problem as a parallel machine problem with multi-modal sequence-dependent setup times as a solving approach. We propose four approaches, a mixed integer linear programming model, a constraint programming model, a constructive heuristic, and a local search heuristic. We will use the simulation tool introduced in the literature to deal with the dynamic arrivals of requests. This simulation tool allows the empirical evaluation of the system's performance and the sensitivity analyses to explore variations in resource requirements to highlight system capacity levels in instances inspired by a real mid-size hospital. The constraint programming model provided a baseline for faster approaches, and it was possible to reinforce the heuristics solutions quality, i.e., the local search approach. Also, from the resources management point of view was possible to identify the number of wheelchairs to meet the system's demands.

  • 11h45 - 12h10

    Healthcare supply network design problems with vendor managed inventory and RFID investment

    • Guoqing Zhang, prés., University of Windsor

    In this paper, we study the healthcare supply location-inventory network design problem with Radio Frequency Identification (RFID) investment, where the sole vendor manages multi-hospital warehouses under a vendor managed inventory (VMI) policy. A robust approach and different algorithms are investigated.